Multi-Disciplinary Senior Design ConferenceKate Gleason College of Engineering

Rochester Institute of TechnologyRochester, New York 14623

Project Number: 11462

THERMOELECTRIC AND FAN SYSTEM FOR COOK STOVE

Jared Rugg / Project Lead Thomas Gorevski / Electrical Team

Bradley Sawyer / Lead Engineer Fahad Masood / Electrical Team

Jeff Bird / Mechanical Team

ABSTRACT

A brief summary (often <200 words) should open the paper. The purposes of an abstract are: (1) to give a clear indication of the objective, scope, and results of the paper so that readers may determine whether the full text will be of particular interest to them; (2) to provide key words and phrases for indexing, abstracting, and retrieval purposes. The abstract should not attempt to condense the whole subject matter into a few words for quick reading. We can work on this after the paper is finished.

NOMENCLATURE

TEGThermocoupleKgMinInCFML – LiterC

INTRODUCTION

Billions of people around the world depend on biomass for everyday cooking fuel. Cooking is often performed on a semi-open stove or over a simple three-stone fire. These methods of cooking are very inefficient and lead to high fuel usage and airborne

pollution. Cooking is often performed indoors and therefore the use of biomass cooking has a negative impact on the health of people who must live in the heavily polluted air conditions. Inefficient cooking methods have especially had a major impact on Haiti, the poorest country in the western hemisphere. Haiti was once covered with forests, but is now less than 4% forested. Much of the deforestation is a result of the need for cooking fuel. The most popular fuel used in Haiti is charcoal.

The introduction of oxygen to any combustion process will, in general, allow for more efficient, complete and clean combustion. The primary function of the thermoelectric and fan system is to provide forced airflow to the combustion chamber of the cook stove without the need of an external power source. A thermoelectric unit is utilized to convert thermal energy taken from the combustion process into electrical energy. Electrical energy provided by the thermoelectric unit is used to power a fan and also a USB port. The USB port can be used to charge the batteries of an ‘auxiliary’ device such as a cell phone.

The thermoelectric generator, which is the heart of the system, develops electrical power when a temperature differential is created between its top and bottom surfaces. The thermoelectric generator (TEG) takes advantage of the Seebeck Effect to create electrical power. The Seebeck Effect happens when a temperature difference is created between two different metals or semiconductors. A voltage differential is created between the metals and when

Copyright © 2008 Rochester Institute of Technology

Proceedings of the Multi-Disciplinary Senior Design Conference Page 2

connected in a loop (circuit) current flows. This configuration is also known as a ‘thermocouple.’ Figure 1 shows how the Seebeck Effect works. Wires constructed of two different metals (metal A, metal B) are joined at two points. When a temperature differential between the two junctions is present (T1, T2) a voltage is created.

FIGURE 1 – SEEBECK EFFECT

A thermoelectric generator consists of multiple “leg pairs” of p-type and n-type semiconductors. Each leg pair is a thermocouple which produces electrical power when a temperature difference is created across it. The leg pairs are connected together in either a series or parallel(?) circuit through thin metal interconnects. The leg pairs are sandwiched between two thin wafers of ceramic giving the thermoelectric generator structural rigidity. The ceramic substrate also provides a smooth contact surface that thermal energy is conducted through. Figure 2 shows a typical thermoelectric generator. This is the style generator that is used in the thermoelectric fan module.

FIGURE 2 – THERMOELECTRIC GENERATOR

The target market for this thermoelectric and fan system is Haiti and, more specifically, a street vendor; therefore, one of the major design objectives is to produce a low cost unit. In conjunction with the low cost, the system should also be simple and easy to use. The beauty of a thermoelectric unit is that it is a solid-state device, meaning it has no moving parts, and is intrinsically simple (from a physical stand-point). Another design requirement is that the unit requires a minimal amount of user actions in order to operate

correctly. User interactions can include adjustment of any knobs, switches or any general physical interaction with the unit. Simple operation of the unit also includes a design that can easily be removed or connected to the stove.

The design of the thermoelectric and fan system has five main facets. The first facet is transferring heat energy from the combustion chamber of the stove to the hot side surface of the TEG. The second facet is removing heat from the cold side surface of the TEG, the third is providing a source of forced airflow to the stove, the fourth is controlling and managing the flow of electrical energy and the fifth is packaging the components in an efficient and simple manner. This thermoelectric and fan project (P11462) is a second generation design as a previous RIT Senior Design team attempted to fulfill the same requirements. The design of the P11462 unit attempts to take the previous team’s design successes and failures into account.

LITERATURE REVIEW

Project 11462 is the second iteration of the thermoelectric and fan system. The previous project (10462) documented much of the important information from their iteration. This year the team attempted to learn from this document information. The main issues faced by P10462 were a malfunctioning electrical system and insufficient cooling of the TEG. The custom heat sink used by team 10462 appears to have insufficient surface area and overall architecture; a fact echoed in their technical write-up. In addition, use of a simple bent metal fan may have provided too low a flow for proper cooling. Another issue that was mentioned was a suspected problem with thermal resistances between the ceramic substrate of the TEG and the surfaces of the thermal bridge (heat conduction rod) and heat sink. Explain previous statement. On the electrical front, team 10426 experienced multiple problems with managing the power allowances between the electrical devices. An important consideration that was discovered through reading the research by previous thermoelectric teams was that the thermoelectric must be operated in a range where it can provide sufficient power. While mainly the shortcomings of team 10462’s design have been addressed, several test methods from their design process were also used. A similar approach is used to better model and empirically determine the heat transfer coefficient of the fire. Also, ideas for determining the flow rate out of the combustion chamber of the stove are used.

Project P11462

Proceedings of the Multi-Disciplinary Senior Design Conference Page 3

PROCESS

The final proposed design was the result of careful consideration of past team experiences, pairwise comparisons of possible solutions, and first principle feasibility calculations. Should discuss the design process in more depth. How did we arrive at these systems? Discuss the high level customer needs and engineering specs. We spent a lot of time on those so they should be mentioned. Also might want to include the systems diagram (saved as tab 2 of most recent block diagram excel file).

The final design consists of a square cross-section metal housing which functions as an air duct. Inside the housing resides a fin style aluminum heat sink along with a computer case fan to provide air flow. Figure 3 shows the design of the thermoelectric and fan system. The air is ducted axially from the top of the stove to the bottom and is then turned 90 degrees to enter the bottom of the stove. This axial design was implemented to fit the system (mainly the large heat sink and fan) in close proximity to the stove. This was desirable for both package-ability purposes and also because keeping the unit close to the stove has a minimal effect on the stove’s center of gravity.

FIGURE 3 – THERMOELECTRIC SYSTEM

One of the most important parts of the design is the heat sink which removes heat from the cold side of the thermoelectric. The simple finned heat sink was chosen because of its simplicity and also because commercially available finned heat sink extrusions are

cheap and plentiful. The finned heat sink also creates a minimal pressure drop when compared to pin style heat sinks. The heat sink is fastened to the metal housing via four cap screws.

The manner in which thermal energy is removed from the combustion process for use by the thermoelectric module was an important point of debate. In the end a “ruler style” steel thermal conduction rod was selected. Although the thermal conductivity of steel is significantly less than that of aluminum, initial tests suggested that the temperatures generated by the fire would cause significant damage to aluminum. The conduction rod enters the side of the stove through a slot and protrudes into the combustion chamber much like a large fin. The other end of the rod transfers energy to the thermoelectric through a small pocket machined in the face of the rod. Compression is maintained on the thermoelectric via four 6-32 cap screws that thread into the heat sink. The thermal bridge that these cap screws create between the conduction rod and the heat sink is an issue. In order to mitigate this problem contact between the threads of the cap screw and the hole in the conduction rod in which they pass through was minimized. Stainless steel washers sit underneath the cap of the screws. A circular recess in the conduction rod secures the washers with a reasonably tight tolerance. These washers serve a two-fold function. Because stainless steel has a relatively low (compared to steel) thermal conductivity the washers help to decrease the thermal bridging through the 6-32 cap crews. Also the washers provide constrain the rod from any lateral motions which could damage the TEG.

The need for a method in which to control the air flow into the stove led to the implementation of the “bypass” design. Initially the air flow into the stove was going to be controlled by varying the voltage to the fan. This idea was abandoned due to the need for constant substantial airflow over the heat sink to provide cooling to the TEG. The bypass is a simple hinged door located downstream of the fan and heat sink. When opened, the bypass allows some air to be ducted to atmosphere instead of into the stove thereby controlling of the strength of the combustion process inside the stove.

The thermoelectric module used for this project is a 4 watt module, meaning that 4 watts is the maximum power output of this module. In order to achieve a 4 watt output the temperature differential across the unit must be approximately 200 C, with a maximum sustained hot side temperature of 300 C. With this data in mind, the unit was designed to operate with a hot side temperature of 300 C and a cold side temperature of 100 C. Most thermoelectric modules, including this one, are approximately 4% efficient. Therefore it was assumed that for a desired output of 4 watts, 100 watts of heat energy would need

Copyright © 2008 Rochester Institute of Technology

Proceedings of the Multi-Disciplinary Senior Design Conference Page 4

to be passed through the TEG unit. A design value of 130 watts was selected to account for thermal losses in the system and for losses in the TEG unit. This value was arbitrary and was a gross estimation. Once these design parameters were selected, the conduction rod and heat sink could be designed to pass 130 watts of energy through the TEG while maintaining a 200C temperature differential across the TEG.FAN

Initially, the proposed maximum nominal specification for air flow was 1.2 kg/min. Using STP, this value converts to ~36 CFM. The flow into the stove will depend on both the resistance to flow of the stove as well as the pressure drop provide by the fan. First, the stove’s impedance needed to be determined; it was decided that this would be determined empirically. A pressure tap was placed in the inlet to the stove. Air was then ducted into the stove with a smooth PVC pipe with a flow meter inline. Using a pressure transducer an outlet voltage was recorded and converted to units of inches of water. Using the known cross-section of the PVC the flow velocity from the flow meter was converted into a flow rate. The flow rate was varied and several data points were taken. The stove’s impedance was then plotted as pressure drop versus flow rate.

0 5 10 15 20 25 30 35 40 450.00

0.10

0.20

0.30

0.40

0.50

0.60Stove Impedence

Flowrate (CFM)

Pres

sure

Dro

p (in

ches

H2O

)

After examining this plot and a general overlay of an estimate of a very efficient computer fan operating at 2 watts it quickly became apparent that the goal of 36 CFM of flow into the stove would not be plausible. Subsequently, a desire to better understand the effect of the air flow rate on the fire’s properties emerged. In order to do this an additional test was run. Again using the PVC with the flow meter inline to duct air into the stove, a fire was started in the stove and the stove was allowed to reach a “hot start” temperature. Next, 2 L of water at room temperature (50 F) was placed in a pot. After the charcoal was topped off to a set level the pot was placed on the stove and the time to boil was measured. This experiment was done at three different flow rates: 20, 30, and 40 CFM. The results are shown in Figure #.

Starting Water Temp (deg F)

Flow Rate (CFM)

Time to boil (min)

51.1 20 3:54

48.6 30 3:15

50.9 40 3:13

As shown in the Figure decreasing the flow rate by half had little effect on the time to boil. With this information in mind, the air flow rate specification was pushed down to a nominal max value of 0.7 kg/min.

The design relies on the fan running at the max flow rate at all times. This is due to the fact that the proper operation of the heat sink relies on good air flow through it. The amount of air entering the fire is controlled by a bypass. Some of the air flow can be ducted out to atmosphere while the rest is ducted into the stove. Currently, this is done through the use of a pivoting door. Because of this design and the engineering specification of being able to produce a range of flow rates into the stove, the fan must be able to produce the maximum flow rate at all times.

A computer case fan was chosen to provide the air flow to the stove. The computer case fan was chosen for a myriad of reasons. First and foremost, computer case fans are cheap and plentiful which fits into the constraint that the system must be cheap and easily producible. An equally important reason is that most computer case fans are relatively efficient and thus require low power to operate. The fan must operate at a low power because power is at a premium in the system. The computer case fan is also desirable because it is well packaged and self-contained which makes it easy to integrate into the system. The life span of a computer case fan is also of importance. A long fan life span is desirable and a computer case fan provides this. The only real issue present with using a computer case fan is that they do not provide large pressure drops.

Once the parts had arrived the flow rate and pressure drop capabilities of the fan needed to be verified. This was done using a test rig designed to specifically test computer case fans. The test rig is a PVC tube with several pressure taps downstream of the fan. The flow of the fan can be varied using a cone at the outlet of the PVC tube. The closer the cone is to the outlet of the tube the greater the resistance to flow and thus higher the pressure drop. The flow rate was measured using a hot wire anemometer that was inserted into the middle of the pipe. Because the fan speed is also a function of input voltage the fan speed was held constant at 12 volts during the test (12 volts is the recommended operating voltage of the fan). The pressure drop was varied and the resulting flow rate recorded. This data was plotted the stove impedance

Project P11462

Proceedings of the Multi-Disciplinary Senior Design Conference Page 5

curve. Despite being one of the most efficient fans available the operating point of the system was still too low. In order to combat this problem the inlet holes to the combustion chamber were enlarged. This modification should decrease the pressure drop into the stove and shift the optimal point to a higher flow rate. This will allow a higher flow rate of air into the combustion chamber of the stove.

ROD

The design of the heat conduction rod involved many factors. To start our process we developed a temperature desired for the end of the rod based on the temperature difference that would provide us with the most power possible. Also an analysis of the thermoelectric module provided us with an approximate heat flux through the module. This piece of information allowed constraining the design to sizes that would allow for that level of heat conduction. A test was also run to determine the U value of the fire (need test name) and a general fire temperature was assumed. The assumptions were used to generate a desired temperature at the inner wall of the stove assuming the section from that point back to the module to be perfectly insulated. So the heat loss from the inner chamber of the stove back was due only to the k value of the material, in this case 1018 steel. The general set of heat transfer equations used for the portion of the rod in the fire was of an adiabatic tip fin. Using these equations a Matlab program was developed to test different lengths and widths for a given thickness to match the amount of heat absorbed by the fire to the desired temperature at the wall of the inner chamber, with that temperature first being calculated with each cross-sectional area and the desired temperature at the thermoelectric. From the different lengths and widths one was picked that was considered the best fit for our design.

HEATSINK

ELECTRICAL

The electrical systems were designed around certain needs by the customer. From an electrical standpoint the priority list is as follows: 1) Provide an airflow into the stove and across the cold side of the TEG, 2) be able to charge a battery pack which will allow the fan to run at startup without TEG power, 3) be able to charge a secondary battery pack that will be able to power the USB device.

BUCK BOOST CONVERTERMultiple circuits were used to accomplish

these three tasks. The design calls for three separate DC-DC buck boost converters. We used the TI

PTN04050 buck-boost DC-DC converter to provide a regulated output voltage. To accomplish this regulated output, a value for RSET was determined for each converter separately. The input of the first converter is being fed directly from the output of the TEG module. At steady-state the TEG can provide a sufficient turn-on voltage for this boost allowing the desired regulated output voltage. This boost has its output connected to two charging circuits and a primary input of a switching circuit. The design of the DC-DC Buck Boost Converter may be seen in the figure below.

This regulated voltage is used as the input for the battery charging circuits. A circuit architecture from the TI corporation is used that allows the design of an output voltage between 5V - 15V from the integrated chip with only a required input voltage of 2.95V - 5.5V. This architecture allows the output of up to 12 W of power. this was a complication we encountered with our first DC-DC buck boost architecture which limited our output power. (Run-on what were you trying to say?)

Both input terminals of the second and third DC-DC buck boost converters are connected to the output node of their respective charging circuit (fan battery charger and auxiliary battery charger). The input of the "auxiliary converter" is connected to the output of the auxiliary charging battery pack which will provide the required turn-on voltage for a USB cable (5V). The concept behind this is to power the USB cable to be able to charge an auxiliary device such as a cell phone or various other devices with a proper hook-up. The input of the "fan converter" is connected to the fan charging battery pack. This converter has its output is then connected to the secondary input of a switching circuit. Both of these converters have been sized to provide the desired output for their respective purposes.

Copyright © 2008 Rochester Institute of Technology

Proceedings of the Multi-Disciplinary Senior Design Conference Page 6

BATTERY CHARGERThe bq2003 charging IC was selected due to

its low power characteristics and because of its versatile charging applications. The charging circuit was configured for a battery pack consisting of four AA cells, a 1.25 mA trickle charge current, and a 1.25 A charging current. The charging circuit also incorporated negative thermal coefficient resistors to track the temperature of the battery pack so that charging would cease under high temperatures as a safety measure. The design of this circuit can be seen in the figure below.

The battery packs were chosen with certain constraints in mind: it had to power the fan at startup and be robust. To satisfy these constraints two battery packs were chosen: one with four AA NiMH cells and the other containing three AA NiMH cells. The chemistry of rechargeable batteries was also taken into consideration. The NiMH chemistry was chosen due to its lack of memory, high energy density, and relative low cost.

NiMH NiCdCapacity (mAh) 1100-3000 500-1000Density (Wh/L) 140-300 50-150

Service Life Up to 1000 Cycles

Up to 750 Cycles

SWITCHING CIRCUITThe LTC4412 switching IC was chosen

because of its logic controlled power switching ability. This architecture is ideal for the design’s application because it is specifically designed to be a low loss power path controller. This circuit is configured to be an automatic switchover of load between two DC inputs with primary and secondary p-channel MOSFET for lowest lost. This allows the higher input supply voltage to be passed through to power the load.This circuit operates by detecting the voltage difference between the Vin and the Sense terminals of the IC. When a secondary source is applied, the Sense pin will be pulled higher than the Vin pin through the external diode. When the secondary power source is greater than the primary source, it completely takes over the stress of the load. The use of external diodes provides protection to prevent current draw to the opposite source. The diagram of this circuit may be seen in the figure below.

The electrical systems were all tested separately to confirm their operation, then as a whole electrical system, and then the entire electrical system was integrated with the mechanical system. The battery charger was tested by monitoring the charge current into the battery pack and also by operating the charger circuit with a voltage source and charging the battery pack from fully discharged to charge in under three hours.

The boost converter was initially tested by itself using a power supply and retrieving data to confirm its operation. Then it was tested by using a test apparatus controlled TEG module to retrieve data which corresponded with the data from the initial test. Then it was tested by using the actual cook stove water boil test to provide power with an uncontrolled TEG module.

The switching circuit was initially tested by using two separate power supplies as inputs to confirm primary and auxiliary functionality using a load resistance. Then a test consisting of a power supply for one input and controlled TEG module for the other

Project P11462

Proceedings of the Multi-Disciplinary Senior Design Conference Page 7

were used to confirm output characteristics and operation. Another test using a controlled TEG module as the auxiliary input and batteries as the primary input was completed and confirmed proper output operation with a load resistance. Does this go into the results and discussion section?

RESULTS AND DISCUSSION

This section should describe your final product or process, whether it met specs (results of testing), and how you evaluated its success. Most conference papers include enough information for your work to be reproducible.

CONCLUSIONS AND RECOMMENDATIONS

This section should include a critical evaluation of project successes and failures, and what you would do differently if you could repeat the project. It’s also important to provide recommendations for future work. Need to discuss ideas for next year. I have some ideas listed in the presentation for tomorrow and I’m sure you guys have some ideas too.

REFERENCES

Within the text, references should be cited in numerical order by order of appearance. The numbered reference should be enclosed in brackets. For example: “It was shown by Prusa [1] that the width of the plume decreases under these conditions.” In the case of two citations, the numbers should be separated by a comma [1,2]. In the case of more than two references, the numbers should be separated by a dash [5-7]. References to original sources should be listed together at the end of the paper, and should include papers, technical reports, books, prior team projects, personal discussions, websites (not Wikipedia), and software. References should be arranged in numerical order according to the sequence of citations within the text. Each reference should include the last name of each author followed by his initials.

(1) References to journal articles and papers in serial publications should include: last name of each author followed by their initials, year, full title of the article in quotes, full name of the publication (abbreviated), volume number (if any) in bold (do not include the abbreviation, "Vol."), issue number (if any) in parentheses (do not include the abbreviation, "No."), inclusive page numbers using “pp.".

(2) Reference to textbooks and monographs should include: last name of each author followed by their initials, year of publication, full title of the publication in italics, publisher, city of publication, inclusive page numbers using "pp.", chapter number (if any) at the end of the citation following the abbreviation, "Chap." (3) Reference to individual conference papers, papers in compiled conference proceedings, or any other collection of works by numerous authors should include: last name of each author followed by their initials, year of publication, full title of the cited paper in quotes, individual paper number (if any), full title of the publication in italics initials followed by last name of editors (if any) followed by the abbreviation, "eds.", city of publication, volume number (if any) in boldface – include, "Vol." if part of larger identifier (e.g., "PVP-Vol. 254") – inclusive page numbers of using "pp.". (4) Reference to theses and technical reports should include: last name of each author followed by their initials, year of publication, full title in quotes, report number (if any), publisher or institution name, city.

Example References:What did we use?Heat transfer bookLast year’s team’s paperAny other papers?[1] Ning, X., and Lovell, M. R., 2002, "On the Sliding Friction Characteristics of Unidirectional Continuous FRP Composites," ASME J. Tribol., 124(1), pp. 5-13.[2] Barnes, M., 2001, "Stresses in Solenoids," J. Appl. Phys., 48(5), pp. 2000–2008.[3] Jones, J., 2000, Contact Mechanics, Cambridge University Press, Cambridge, UK, Chap. 6.[4] Lee, Y., Korpela, S. A., and Horne, R. N., 1982, "Structure of Multi-Cellular Natural Convection in a Tall Vertical Annulus," Proc. 7th International Heat Transfer Conference, U. Grigul et al., eds., Hemisphere, Washington, DC, 2, pp. 221–226.[5] Hashish, M., 2000, "600 MPa Waterjet Technology Development," High Pressure Technology, PVP-Vol. 406, pp. 135-140. [5] Watson, D. W., 1997, "Thermodynamic Analysis," ASME Paper No. 97-GT-288.

[6] Tung, C. Y., 1982, "Evaporative Heat Transfer in the Contact Line of a Mixture," Ph.D. thesis, Rensselaer Polytechnic Institute, Troy, NY.[7] Kwon, O. K., and Pletcher, R. H., 1981, "Prediction of the Incompressible Flow Over A Rearward-Facing Step," Technical Report No. HTL-26, CFD-4, Iowa State Univ., Ames, IA.

ACKNOWLEDGMENTS

Copyright © 2008 Rochester Institute of Technology

Proceedings of the Multi-Disciplinary Senior Design Conference Page 8

Be sure to acknowledge your sponsor and customer as well as other individuals who have significantly helped your team throughout the project. Acknowledgments may be made to individuals or institutions.Who to thank: Edward Hanzlik, Robert Stevens, John Wellin, Hoople, Rob Kraynik

Project P11462